Lead/Presenter: Sarah Ahmed, COIN - Houston
All Authors: Shah NR (Providence VAMC, Dept. of Medicine & Dept. of Health Services, Policy & Practice, Brown University Warren Alpert Medical School, Providence, RI), Winchester DE (Malcom Randall Gainesville VA Medical Center, Department of Medicine, University of Florida College of Medicine, Gainesville, FL), Ahmed ST (Health Policy, Quality, & Informatics Program, Center for Innovations in Quality, Effectiveness & Safety, Michael E. DeBakey Houston VA Medical Center, Section of Health Services Research, Department of Medicine, Baylor College of Medicine, Houston, TX) Ramsey DJ (Health Policy, Quality, & Informatics Program, Center for Innovations in Quality, Effectiveness & Safety, Michael E. DeBakey Houston VA Medical Center, Section of Health Services Research, Department of Medicine, Baylor College of Medicine, Houston, TX) Akeroyd JM (Health Policy, Quality, & Informatics Program, Center for Innovations in Quality, Effectiveness & Safety, Michael E. DeBakey Houston VA Medical Center, Section of Health Services Research, Department of Medicine, Baylor College of Medicine, Houston, TX) Wu W (Providence VA Medical Center, Department of Medicine, Brown University Warren Alpert Medical School, Providence, RI) Waldo SW (VA Eastern Colorado Health Care System, Department of Medicine, University of Colorado School of Medicine, Denver, CO) Schofield RS (Malcom Randall Gainesville VA Medical Center, Department of Medicine, University of Florida College of Medicine, Gainesville, FL) Ballantyne, CM (Section of Cardiovascular Research, Department of Medicine, Baylor College of Medicine, Houston, TX) Petersen LA (Health Policy, Quality, & Informatics Program, Center for Innovations in Quality, Effectiveness & Safety, Michael E. DeBakey Houston VA Medical Center, Section of Health Services Research, Department of Medicine, Baylor College of Medicine, Houston, TX) Virani SS (Health Policy, Quality, & Informatics Program, Center for Innovations in Quality, Effectiveness & Safety, Michael E. DeBakey Houston VA Medical Center, Section of Health Services Research, Department of Medicine, Baylor College of Medicine, Houston, TX; From the Section of Cardiovascular Research, Department of Medicine, Baylor College of Medicine, Houston, TX)
Objectives:
We aimed to quantify facility-level variation in cardiac stress test utilization in a national cohort of patients with ischemic heart disease.
Methods:
Veterans Health Administration (VA) datasets were used to retrospectively identify patients with ischemic heart disease (IHD) and an outpatient primary care clinic visit between October 2013-September 2014. Facility-level overall cardiac stress test utilization, defined as the total number of cardiac stress tests (exercise treadmill tests [ETT], stress echocardiograms [SE], SPECT stress myocardial perfusion imaging [MPI] and PET stress MPI) performed/100 IHD patients/year, was calculated for 130 VA medical centers. Non-imaging (ETT) vs. imaging-based stress test utilization rates were also calculated. For overall stress tests, ETT, and imaging-based stress tests, multivariable 2-level hierarchical regression was used to determine the unadjusted median rate ratio (MRR), a measure of the likelihood that two random facilities differ in test utilization for identical patients. MRR was then recalculated adjusting for age, gender, race, hypertension, diabetes, and diagnostic cost group relative risk score (a surrogate measure of overall illness burden).
Results:
Unadjusted facility-level cardiac stress test utilization markedly varied, ranging from 2.1 studies/100 IHD patients to 31.7 studies/100 IHD patients with a mean (SD) of 15.4 (5.5) studies/100 IHD patients. The unadjusted MRR for overall cardiac stress test utilization was 1.69 (95% confidence interval [CI]: 1.58-1.79). Adjusting for patient-level characteristics including rurality, the MRR decreased to 1.40 (95% CI: 1.34-1.46). The unadjusted MRR for ETTs was 1.69 (95% CI: 1.58-1.79) and was 1.50 (95% CI: 1.42-1.57) after patient-level adjustment. The unadjusted MRR for imaging-based cardiac stress tests (SE, SPECT MPI and PET MPI) was 1.31 (95% CI: 1.27-1.35) and was 1.44 (95% CI: 1.37-1.50) after patient-level adjustment. County-level geographic variation in overall stress test utilization did not appear to be significantly correlated with rurality.
Implications:
Our finding of 40% residual facility-level variation in cardiac stress test utilization after adjustment for patient-level characteristics within the VA represents a potential opportunity to improve efficiencies in the care of Veterans with ischemic heart disease.
Impacts:
To improve efficiencies by identifying variation in cardiac stress test utilization by facility in the care of veterans with ischemic heart disease